J. Phys. Chem. 1993,97, 13477-13483
13477
Intramolecular Dynamics and Unimolecular Decomposition of Polyatomic Molecules in Rare-Gas Clusters: Cluster Size Effects in H202-Ar~(N = 1-39) Lisa M. Finney and Craig C. Martens' Department of Chemistry, University of California, Imine. Imine, California 92717-2025 Received: August 2, 1993"
The cluster size and structure dependence of local microsolvation environment on intramolecular energy transfer and unimolecular dissociation is studied for the systems H202-Ar~ (N = 1-39) using the method of classical trajectory simulation. The rates and mechanisms of intramolecular energy transfer and molecular decomposition of the complex are investigated as a function of N. Three major mechanisms for cluster modification of isolated molecule behavior are operative: vibrational deactivation of the excited molecule, modification of intramolecular energy redistribution, and cluster-induced recombination. The evolution of these effects is studied as cluster size and structure are varied. It is found that the degree of solvation of the hydrogen peroxide molecule, rather than the absolute cluster size, determines the extent of the cluster effects.
I. Introduction The propertiesof weakly-bound clusters are currently the focus of considerable experimental and theoretical Finite clusters constitute a novel state of matter which is, in some sense, intermediate between isolated few-body systems and true condensed phases. Dynamical processes in clusters provide models of solvation,'O-21 solvent caging,'O~~~-~~ phase transitions,3347 d i f f ~ s i o n , ' ~chemical * ~ ~ - ~ ~reactions10~13-16,22,23,26,52-67 and other condensed-phase processes in well-defined systems where the rigorous methods of gas-phase chemical physics can be applied. By systematic variation of the size of clusters in an experiment or simulation, the evolution of behavior from the limits of small complexes to the condensed phase can be investigated. In this paper, we study the cluster size dependence of the intramolecular energy redistribution and unimolecular decomposition dynamics of a polyatomic molecule associated with a rare-gas cluster. The particular systems considered are H202ArN, for N = 1-39. We apply the method of classical trajectory simulation to model the dynamics following overtone excitation of an OH stretch for a series of cluster sizes and compare the results with analogous calculations on the isolated molecule. In a previous publicati~n,~~ we described the dynamics of the particular case of H20~Ar13.Here, we extend that study and investigatethedependenceof the rates and dynamical mechanisms of intramolecular energy redistribution and unimolecular dissociation on the size and structure of the Ar solvent cluster. Size effects in clusters have been the focus of a wide range of experimentaland theoretical investigation^.^-^,^^-^^ These studies have addressed a number of aspects of the chemistry and physics of finitecluster systems,including the emergenceof bulkstructure with increasing the size dependence of electronic ~tructure,~ the role of quantum effect^,^^^,^^ finite-size thermodynamics,34~35~38~75-76 and influence ofsize and structure on cluster s p e c t r o ~ c o p y ~and ~.~~*~~ Issues such as cluster rigidity, isomerization, and finite system analogues of phase transitions have also been a d d r e s ~ e d . ~ ~ , ~ ~ The size dependence of molecular dynamics in highly excited clusters has been the focus of a number of experimental and theoretical studies. Amar and Berne investigated the role of solvent caging on photodissociation and recombination in moleculardynamics simulationsofthe BrrArNsystems ( N = 8-70).23 Theoretical studies of the photodissociation dynamics of Br2- in Ar and COz clusters of varying size have been reported;s0 this
* To whom correspondence should be addressed. *Abstract published in Advance ACS Absrrucrs. November 15, 1993. 0022-365419312097-13477$04.00/0
work was motivated by mass-resolved experiments performed on Br2- in C02 clusters by the Lineberger group.22 Photofragmentation studies of 12- in small ( C 0 2 ) ~clusters (N = &22) by Lineberger and co-workers have focused on the effect of cluster size on photodissociation and recombinant dynamic^.^^^^^ Monte Carlo calculations by the same and molecular dynamics simulations by Amar and Perera on I ~ - ( C 0 2 ) 2have ~ provided a theoretical perspective of size-dependent caging effects in clusters. Molecular dynamics simulations by Perera and Amar involving small molecules associated with van der Waals clusters have addressed the dependence of spectral shifts on cluster size and structure;82 these theoretical studies were motivated by infrared spectroscopic investigations of Scoles and co-workers of molecules in Ar ~ 1 ~ s t e r s . ~ ~ ~ ~ ~ , ~ ~ - ~ ~ In ref 53 we presented a computational study of the dynamics of overtone-induced unimolecular decomposition of hydrogen peroxide associated with an Aril cluster. We identified three major effects of cluster solvation on the dynamics of the H202 molecule: (i) vibrational deactivation of the excited moleculevia both molecule to cluster energy relaxation and cluster-mediated intramolecular vibration-rotation energy transfer, (ii) cluster modification of intramolecular vibrational energy redistribution, and (iii) cluster-induced recombination of the OH fragments. In the present study, we examined the evolution of these effects with both the size and the structure of the ArN cluster microsolvation environment. Classical trajectory methods, combined with strictly classical initial conditions, are employed to model the dynamics. Cluster sizes with N = 1-39 are considered, and the results are compared with those of the isolated molecule. For geometries with the H202 molecule in a highly solvated site (Le., "inside" of the cluster),we investigatethedependenceof dynamical rates and mechanisms of the number of cluster atoms. For a fixed N = 13, we assess the effects of changing the geometric structure of the cluster, comparing H202 at the center of the cluster with H202 in a cluster surface site. The size and structure dependences of the recombinationprobabilityare also investigated. As described below, we find that it is the degree of solvation of the molecule by the cluster atoms, rather than the absolute size of the cluster, that determines the extent of the dynamical effects of microsolvation. The dynamics of isolated HzOz in the gas phase has been the subject of recent e~perimental*~-~~ and t h e o r e t i ~ a studies. l~~~~~ We also note recent and related work on overtone-induced decomposition of HOC1 in Ar gas and liquid reported by Li et a1.103 The organization of the rest of this paper is as follows: In 0 1993 American Chemical Societv
Finney and Martens
13478 The Journal of Physical Chemistry, Vol. 97, No. 51,1993
section 11, the classical trajectory method used to simulate the dynamics of H20rArN is described, and the potential energy surface employed in this study is presented. In addition, the methodology used to find energy minimum structures, generate initial conditions, integrate classical trajectories, and analyze the resulting data is described. The results of the simulations are presented in section 111, along with a discussion of the observed size- and structure-dependent trends in the dynamical behavior. Finally, a summary is given in section IV.
n
N=l
N=2
N=lO
N=13
N=7
11. Metbod
The dynamics of isolated H202 and the molecule-cluster complexes H 2 0 r A r ~are modeled by classical trajectory simulation. In this section, we describe the potential energy surface and outline the methods employed to generate minimum energy geometriesand trajectory ensemble initial conditions. In addition, we outline our analysis of the resulting simulation data. (A) Potential Surface. The potential energy surface of the H 2 0 r A r ~system used in this study is of the general form
v = vm+ v,+ vm4
N=13
(surface)
(1)
The first term, V m , is the intramolecular potential energy surface of the isolated hydrogen peroxide molecule, while V,and Vm-c are Ar-Ar and Ar-H202 interactions,respectively. For J",we adopt the potential surface developed by Getino et al., which is based on a fit to ab initio calculations.lo2 The intramolecular potential consists of three terms: (2) = vstretch + vbend + Torsion corresponding to two-body stretches, three-body bends, and fourbody torsional interactions, respectively. The two-body stretch potential is modeled by a sum of Morse oscillators: vm
N-26
N=39
Figure 1. Minimum geometry structures of H 2 0 2 - A r ~clusters.
TABLE I: Local Minima Energies system
energy (cm-l) -157.21 -383.54 -1958.93 -2985.73 -4284.14 -9290.40 -1 5261.86
3
(3) where Di are the bond dissociationenergies, and pi are scale factors. The ri and ri' are instantaneous and equilibrium bond distances, respectively. The three-body term corresponds to the OOH bend degrees of freedom and is given as the sum of harmonic interactions, each with a force constant k:
H20rAr13 (surface)
-4151.84
calculations for the isolated molecule.lo2 The parameters used in V m are listed in Table 111, and the resulting normal-mode frequencies are given in Table 11. See refs 53 and 102 for further details of the intramolecular potential function. The cluster interactions, V,and J"+,are modeled by pairwise additive potentials. Each atom-atom interaction is taken to be a Lennard-Jones potential function:
Here 8i and 8ieare the instantaneous and equilibrium bend angles, respectively. The force constants are attenuated by switching functions S(r),which we describe below. The four-body torsion is represented as a trigonometric series: N,
3
N,
m=O
where 4 is the torsion angle, and Am are constant Fourier components. The torsional potential is also modified by switching functions. Switching functions are used to model the attenuation of the bend and torsional interaction strengths with OH and 00 bond lengths. The switching function used for the OH bonds is given by
The corresponding 00 bond switching function is taken to be
Si(ri)= 1 - tanh[C(r, - r:)']
(7)
These switching functions were determined from fits of ab initio
where e,, eu, a,, and u~ are the appropriate Lennard-Jones parameters. Nc and N m are the number of cluster and molecule atoms, respectively. See Table I11 for numerical values of the potential parameters. (B) Cluster Geometries. Local minimum energy structures for each cluster size were found by kinetic energy quenches of high energy trajectories,which were started from estimated initial geometries. Several heating and cooling cycles were performed in an attempt to find the structures corresponding to global energy minima. The lowest energy geometries found for each molecule cluster system are illustrated in Figure I, and the energies corresponding to these structures are given in Table I. The lowenergy structures for clusters with 13 or more Ar atoms are characterized by a complete "solvation shell" surrounding the H202 molecule. Other local minima, with the hydrogen peroxide
The Journal of Physical Chemistry, Vol. 97, No. 51, 1993 13479
Polyatomic Molecules in Rare-Gas Clusters
TABLE II: Normal-Mode Frequencies' mode OH sym OH str OOH sym bend
w
(cm-I) 3883 3882 1474 1402 406
OOH asym bend torsion Normal-mode frequencies of hydrogen peroxide potential, eq 2. TABLE III: Potential Parameters 88.6 kcal/mol 0.965 A 2.62.A-I 49.6 kcal/mol 1.452 A 2.58 A-I 1 .I45 rad 0.99 mdyn A/rad2 0.865 kcal/mol 3.453 kcal/mol
3.206 kcal/mol -0.488 kcal/mol 0.0638 kcal/mol 3.410 A 0.1706 kcal/mol 3.180 A 0.2379 kcal/mol 3.110A 0.1205 A-5 1.58 A-2
*adsorbed"on the surface of the cluster, were found for the larger systems. However, these surface structures are higher in energy than their highly solvated isomers. For general N, we considered only the lowest energy structure. Calculations were performed using both the lower energy 'caged" structure and the higher energy *adsorbed" geometry for H~Oz-Ar13. The structure of the N = 13 surface site cluster is also shown in Figure 1, and its energy is included in Table I. (C) Initial Conditions. The minimum energy geometries described above were used to generate ensembles of initial conditions for the trajectory calculations. For a given configuration, each atom of the system was assigned a random velocity from a Maxwellian distribution corresponding to a system temperature of 10 K. The velocities were then adjusted to yield zero total linear and angular momenta. The ensemble of initial conditions was then propagated for 120 ps, with periodic rescaling of the velocities to correct the time-averaged kinetic energy to yield the desired initial temperature of 10 K. For a given N, ensembles of 200-300 trajectories were generated by repeating this process for each trajectory, starting with different initial random velocities. The thermal ensemble was then subjected to overtone excitation of an OH local mode vibration. This was accomplished by adjusting the OH bond momentum of one OH stretch, chosen randomly, to yield a classicalenergy corresponding to the desired vibrational quantum number u. The positions and momenta of all atoms were then stored and used as initial conditions for the trajectory simulations. The initial conditions employed in this study are purely classical, in that they do not take into account zero-point energy in the vibrational modes of the cluster or molecule. This is necessary to avoid spurious dynamical processes, which can be induced by redistributionof the classical zero-point energy. For a given initial overtone excitation, the absence zero-point energy in the other vibrational modes results in a molecule with total energy significantly less than the case where it is included. Thus, the unimolecular rate calculated here will be less, for a given u, than the result of calculations performed on the same potential surface which include the zero-point energy. Classical trajectory calculationsperformed on this surface with zero point energy included are reported in ref 102 and are in general qualitative agreement with experiment. We have chosen to neglect zero-point energy and treat all processes as purely classical. Consequently, this system must be viewed as a model problem and is not meant to be a realistic representation of the quantum dynamics of a hydrogen peroxide molecule. (D) Classical Trajectory Integration. Ensembles of classical trajectories were propagated for 33 ps by numberical integration of Hamilton's equations. A sixth-order hybrid Gear routine'"
was employed for the numerical integrations. The accuracy of these long integrations was monitored by energy conservation, as well as by conservation of the total linear and angular momenta. Errors in the total energy were approximately 1 part in 1 0 000. (E) Trajectory Data Analysis. The ensemble-averaged energy transfer within the cluster-molecule system was monitored as a function of time. The total energy was separated into molecule, cluster, and molecule-cluster interaction contributions. The energy of the H202 molecule was partitioned into vibrational, rotational, and translational energies, and the vibrational energy was further divided (approximately) into the six internal modes using the method of Wilson, as described in ref 102. The dynamiical G matrix was calculated as a function of time, and the diagonal elements used along with the single-mode potential functions to calculate the mode energies:
where r(t) and p ( t ) are the instantaneous values of the internal coordinates and conjugate momenta, respectively, and 6 is the term in the potential corresponding to modej. For the bend and torsion, the stretch-dependent switching functions are inciuded in the definition of 6.This method neglects the kinetic cross terms coupling the internal degrees of freedom, which make a relatively small contribution to the vibrational enzrgy, especially after ensemble averaging is performed. The time histories of these energies were stored for each member of the ensemble, in addition to their ensemble-averaged values. The unimolecular decomposition times were determined for each trajectory by monitoring the 00 bond distance as a function of time and using a critical extension as a condition for dissociation. The actual time of dissociation was taken to be the time of the previous inner turning point of the 00 bond. We also employed an energy criterion, where dissociation was taken as the time when the 00 bond energy exceeded DOO. Recombination was assumed to have occurred when this mode energy dropped back below Dm. Both methods gave equivalent results for dissociation. The energy criterion was more convenient for detecting recombination events.
III. Results and Discussion In this section,we describe the results of trajectory calculations of HzOrArNdynamics for severalN, and for two levels of initial OH vibrationalexcitation: u = 5 and u = 7. Foru = 5 , dissociation cannot occur for our system, while decomposition is energetically possible for u = 7. The results obtained for molecule-argon cluster complexes are compared with those for the isolated molecule using analogous classical initial conditions. The smallest cluster system studied consists of the HzO2 molecule bound to a single Ar atom, while the largest system is HzOz-Arjg. Several intermediate N are also considered, so that the complete study treats N = 1, 2, 7, 10, 13, 26, and 39. The isolated system is also treated for comparison. For the N = 13 system, twogeometries areconsidered: a cluster with the molecule highly solvated by thecluster atoms, and a distinct (higher energy) geometry with the molecule on the surface of the cluster. The structures of these systems are displayed in Figure 1. we investigated the dynamics of the N In a previous = 13 system. Three important dynamical mechanisms affecting the overall behavior were identified: (i) vibrational deactivation of the excited molecule, (ii) cluster modification of intramolecular energy-transfer dynamics, and (iii) cluster-induced recombination of nascent OH fragments. In the present work, we examine the dependence of these processes on both the size and the initial geometry of the cluster-molecule complex. Vibrational Deactivation. In ref 53, it was observed that the Ar cluster induces vibrational energy relaxation of the hydrogen peroxide molecule in H202-Arl3, which can lead to deactivation of potentially dissociative trajectories. The vibrational deacti-
13480 The Journal of Physical Chemistry, Vol. 97, No. 51, 1993 8000
7500 7000
6500 6000 5500
5000
0
5
10
15
20
25
30
35
40
Time (ps)
a 0.006 0.005
2
0.002 0.001
b
t
// ,
K
2
N=13
surface
N
Figure 2. Vibrational deactivation of hydrogen peroxide by the argon
cluster. The curves shown are ensemble averaged vibrational energies of H202 embedded in clusters Arl, Ar2, Ar7, Arlo, Ar13.Ar26, Ar39 for u = 5. Also plotted is the ensemble averaged vibrational energy of HzO2 adsorbed to the surface of an Arl3 cluster.
vation was found to occur by two mechanisms: (i) moleculecluster energy transfer in the form of vibrational and rotational relaxation and (ii) cluster %atalyzed’’ energy flow from internal vibration to molecular rotation. Here, we investigate the dependence of vibrational energy relaxation on the size and structure of the cluster. For this portion of the study, we consider ensembles with initial OH excitation of u = 5; the resulting energy is below the dissociation threshold of our model. In Figure 2, we show resultsfor thevibrational energy relaxation of the isolated molecule and the N = 1,2, 7, 10, 13,26, and 39 highly solvated geometries. In addition, the N = 13 surface geometry data is displayed [labeled An&)]. Figure 2 (a) gives a composite plot of the vibrational energies of each ensemble vs time. The total vibrational energy of H202 was approximated by the sum of the internal mode energies, as given in eq 10, and averaged over each ensemble. Figure 2b shows the vibrational relaxation rate of each ensemble as a function of N, determined by fitting the logarithm of the data shown in Figure 2a for times between 10 and 25 p;the decay was reasonably well-represented by an exponential in this time range for all ensembles. Figure 2 demonstrates that the rate of vibrational energy deactivation depends on both the size and the structure of the cluster. For the highly solvated series, the vibrational relaxation rate increases monotonically with cluster size. The isolated molecule has a conserved vibrational energy. Adding a single Ar atom to the system causes slow energy flow out of the vibrational degrees of freedom. Additional cluster atoms lead to increases in the rate of vibrational deactivation. The largest changes in behavior occur as the first solvation shell around the H202 molecule is built up and completed at N = 13. For N > 13, the addition of atoms to the cluster leads to only small changes in the relaxation behavior. The N = 13 surface site ensemble exhibits a vibrational relaxation rate that is slightly less than the N = 7 ensemble. Inspection of Figure 1 shows that the extent of HzOz solvation
Finney and Martens is similar in these two cases and demonstrates that the structure of the molecule solvation environment is the most important characteristic of the system influencing the vibrational relaxation dynamics, rather than the absolute size of the cluster. Cluster Modification of Intramolecular Vibrational Energy Transfer. In ref 53, we described the effect of cluster solvation OII the intramolecular vibrational energy redistribution (IVR) dynamics of H20rAr13. Here we investigate the influence of cluster size and structure on IVR. We again consider the u = 5 ensemble, to avoid the competing process of unimolecular decomposition. In parts a 4 of Figure 3 we show the ensemble-averaged mode energies for the isolated molecule and the N = 1, N = 7, and N = 13 ensembles, respectively. The data are labeled as follows: OH1 is the energy of the initially excited OH stretch; OH2 is the energy of the OH stretch that is unexcited initially; B1 is the energy of the bend adjacent to the initially excited OH stretch; B.! is the energy of the distant OOH bend; 00 is the energy of the 00 bond, and TOR is the torsional energy. These quantities ware evaluated by averaging the energies of each mode over the trajectory ensemble. Mode energies were evaluated as described in section 1I.E. For each ensemble, the OH1 initial energy is approximately 17 700 cm-I. The scale of the plot has been chosen to emphasize lower energies, so the initialOH 1excitationis off-scale. Following initial excitation, the excited OH stretch relaxes into the other modes of the molecule. It should be noted that equipartition of the energy is not complete by the end of the 33-p integration. As thesizeofthecluster increases, the behavior ofthemolecular mode energies evolves from the isolateddynamics shown in Figure 3a to the dynamics of the completely solvated molecule, given in Figure 3d. For larger cluster sizes (not shown), the results are similar to those for the N = 13 ensemble. A number of sizedependent trends can be observed. First, the OH1 relaxation rate increases with increased solvation. The initial step of OH stretch relaxation is energy transfer into the bend, and this can be promoted by collisions between the vibrating H atom and the solvent Ar atoms when the cluster is present. A second effect is inhibitionof energy flow into the internal rotation of the molecule as the cluster size increases. Although this low-frequency mode does not couple strongly to the higher frequency stretches and bends, causing energy transfer into TOR to lag behind that into other modes, nearly 1000 cm-’ of torsional excitation is present on average at the end of the 33-ps simulation for the isolated H202 ensemble. As the solvation of the molecule is built up, the increased intermolecular interactions inhibit internal rotation, leading to less than 200 cm-l of energy in TOR for the N = 13 ensemble at 33 ps. The cluster effect on torsional motion is further illustrated in Figure 4. Here, the ensemble-averaged torsional energy vs time is presented for each of the ensembles. For the highly solvated geometries, the extent of energy flow into TOR decreases monotonically with cluster size. The ensemble with hydrogen peroxide on the surface of an N = 13 cluster shows behavior that is quite similar to the N = 7 ensemble. Again, this results from the similar local solvation environment of H202 in the two cases. The most significant modification of IVR dynamics involves the extent of energy equipartition between the bends and the 00 stretch. For the isolated molecule ensemble, shown in Figure 3a, the energies of the B1, B2, and 00 subsystem have equilibrated by the end of the simulation. The initial energy relaxation step is transfer from OH1 into B1. The two bend modes can rapidly exchange energy, and the rises of the B1 and B2 energies seen in Figure 3a are very similar. The energy transfer into the 00 stretch is somewhat slower than bend-bend exchange, leading to a lag in the 00 energy. By approximately r = 20 p, however, nearly complete redistribution of available energy has occurred between B1, B2, and 00.
Polyatomic Molecules in Rare-Gas Clusters
The Journal of Physical Chemistry, Vol. 97, No. 51, 1993 13481
8000
8000
7000
7000
n6000
76000
U
!$ 5000
5000
W
4000
3
i3 4000 8 3000
3000
wG
2000
2000
1000
1000
0
0 10
5
0
a
20 2 5
15
30 35 4 0
Time (ps)
0
C
5
10 15 20 25
30 35 4 0
Time (ps)
8 8000 8 h 3
'5
1
"'I
6000 5000
W
5
w
4000 3000 2000
4
n
6000
yd
5
v
5000
&
4000
w
2000
3000 1000
1000
0
0 0
10
5
b
20 2 5
15
30 35 4 0
Time (ps)
0
d
5
10
15
20 2 5
30 35 4 0
Time (ps)
Figure 3. Ensemble-averaged mode energies for (a) isolated Hz02 and H202 associated with (b) Arl, (c) Ar7, and (d) Arl3. Each ensemble has an initial OH excitation corresponding to u = 5. The labeling is as follows: OH1 is the initially excited OH bond energy, OH2 is the other OH bond energy, B1 is the energy of the bcnd mode adjacent to the initially excited OH stretch, B2 is the energy of the distant bend, and TOR is the torsional energy. The vertical pition of these labcls indicate the ordering of the magnitudes of the mode energies. 12001
t
n
Y
sz,
w
'
I
'
I
,
I
'
I
'
I
'
I
'
I
,
3000
\
n
-6
loo0 800
W
ti5
I
2500 2000
v
W
1500
400
w= 1000 'um 500
200
0
6oo
Isolate
-500
0 0
10
20
30
40
Time (ps)
0
5
1C
15
20 25
30 35 40
Time (ps)
Figure 4. Ensemble-averaged torsional (TOR) energies of the H 2 0 ~ ArNsystems with initial excitation corresponding to u = 5. The size and structure of the Ar clusters are labeled.
Figure 5. Difference between the two ensemble averaged bend energies (Em - E82) for the molecule-cluster complexes with initial excitation corresponding to u = 5, for all the value of N considered.
As solvent atoms are added to the system, the character of the (Bl, B2, 00) subsystem changes. Increased solvation of the molecule leads to a breakdown of complete energy equilibration between these modes. For N = 1, the dynamics closely resemble the isolated molecule case. As the size of the cluster increases to N = 7, however, it is found that the energies of B1 and B2 have not equilibrated by the end of the simulation. The difference in bend energies becomes even more pronounced for the N = 13 results shown in Figure 3d. Here, the bend adjacent to the initially excited OH stretch, B 1, has nearly twice the energy of the distant bendat t = 33ps. TheOOenergyoftheensembleisapproximately the average of B1 and B2 energies. In Figure 5 we examine systematically the "disequilibrium" of bend energies as a function of cluster size and structure. Here, we show a composite plot of Eel - E Bvs~ time for all ensembles considered. At early times, all ensembles show a short-time positive value for this quantity, as energy is transferred from OH 1 into B 1. This transient quickly decays as the bends exchange
energy for the isolated, N = 2, N = 7, and N = 13 (surface) ensembles, which are characterized by incomplete solvation of the H202 molecule. As the first solvation shell becomes nearly complete for N = 10, a transition in behavior occurs, with the initial disequilibrium of bend energies persisting at longer times. The behavior of the ensembles with complete first solvation shells (N = 13, 26,and 39) are all similar, with a positive deviation of bend energies of roughly 2000 cm-I existing at the end of the simulation. The cluster environment is clearly affecting the intramolecular dynamics of hydrogen peroxide. As the degree of solvation by cluster atoms increases, large-amplitude molecular vibrations become increasingly inhibited by molecule-cluster repulsive interactions. Both torsional excitation and facile bend-bend energy transfer are suppressed by the presence of the associated Ar cluster. Unimolecular Dissociation and Recombination. We now consider ensembles of trajectories with initial OH excitation
Finney and Martens
13482 The Journal of Physical Chemistry, Vol. 97, No. 51, 1993 200
1''
L
160
&
"
"
"
'
I
"
' / I
"
2
120
' -
-. -
5
.d
3
"
.*
W
+
'',:,"
-
ao
-
N=13 surface' 40
1
0
" ' " " " ' ~ " " ~ " " " " ~ J
0
5
10
15
20
25
N Figure 6. Unimolecular lifetime of H202as a function of cluster size and structure, for an initial OH excitation of v = 7. See the text for further details.
0 0
5
10
15
20
25
30
N Figure 7. Cluster size dependence of recombination probability for molecules with an initial OH excitation of u = 7.
correspondingtou = 7. In this case, themoleculescontain enough energy to undergo unimolecular dissociation. For the isolated molecule with classical initial conditions (i.e., no zerepoint energy added to thevibrationaldegreesof freedom), thislevelofexcitation results in a unimolecular lifetime of approximately 26 ps.53 As cluster atoms are added to the system, the unimolecular kinetics and dynamics are modified, altering the lifetime of the overtone excited hydrogen peroxide molecule. The dependenceof the unimolecular lifetime of H202(u=7)on cluster size is shown as the solid curve in Figure 6. The lifetime increaseswith Ngradually from the isolated molecule value until N = 7. Beyond this cluster size, the lifetime lengthens dramatically, becoming too long to determine from our simulations for N > 13. The slowing of the dissociation rate with N is due to a combinationof effects: vibrational deactivation, modified IVR, and recombination events. A detaileddiscussionof these processes for the case of N = 13 was given in ref 53. One effect of the cluster on the unimolecular decomposition of H2Oz is its ability to cause recombination of the nascent OH radicals. The atoms of the cluster provide a channel for translational and rotational energy relaxation of thedecomposition products and exert attractive van der Waals forces which can bind the fragments to the surface of the parent cluster. Both of these processes become more effective as the degree of solvation of the initial molecule is increased, and recombination is a major component of the increasing lifetime with cluster size. Figure 7 shows the probability of recombination as a function of N for the u = 7 ensembles. The recombination probability increases with increasing cluster size and is approximately 80% for H20~Ar26. The solid curve of Figure 6 is based on the kinetics of the decay of the total number of intact hydrogen peroxide molecules .with time. If recombination events can occur, the effective lifetime of the moleculeis lengthened. To investigatethe size dependence of the cluster on unimolecular decomposition beyond the effect of recombination, we calculated an alternative lifetime which
characterizes the decompositionof H202 irrespective of whether later recombination occurs or not. For the determination of this quantity, shown as a dashed curve labeled T I on Figure 6,a molecule is removed from the ensemble once it undergoes its initial decomposition. This lifetime thus is not influenced by subsequentrecombination. The growing differencebetweenthese two types of lifetimes with N indicates the increasing dominance of recombination in the unimolecular decompositiondynamics of the peroxide molecule with cluster size. Also shown in Figure 6 are the data for H~02-Ar13in the surfaoc site geometry. Again, the behavior of this adsorbed geometry shows less deviation from the isolated molecule case than the highly solvated N = 13 (surface) ensemble and is close in behavior to the N -5 7 case, where the local microsolvation structure is similar.
Iv. summary In this paper, we have presented a systematic study of the size and structure dependence of intramolecular energy transfer and unimolecular decomposition of HzOrAr, following initial OH overtone excitation. This work extends the study reported in ref 53, where the single case of N = 13 was investigated, to consider the evolution of dynamical behavior with the number of Ar atoms and structural variations of the local microsolvationenvironment of the HzOz molecule. Three major processes have been explored: (i) vibrational deactivation of the initially excited molecule, (ii) modificationof the rate and extent of intramolecular energy transfer, and (iii) cluster-inducedrecombinationof nascent OH fragments, and the effect of these processes on the H202 unimolecular dissociation dynamics has been characterized as a function of cluster size and structure. For a series of clusters with minimum energy geometries corresponding to the H2Oz in a highly solvated site, the effects of the cluster become more pronounced with size, and show the most dramatic changes as the first solvation shell around the HzOz is completed. For a fixed cluster size, the magnitude of the cluster's influence depends on the degree of solvation, with a highly solvated molecule showing much greater modification of its behavior than a molecule on the surface of a cluster of the same size. From the perspective of observable rates of reactions, the major effect of the cluster is a lengtheningof the lifetime with increasing solvation of H202 by argon atoms. The particular systems considered here illustrate the general evolution of solvent effects on IVR and unimolecular reaction, from the isolated molecule limit, through small complexes of a few components, and finally to relatively large clustes containing tens of atoms. The study thus provides a bridge between previous contributions on chemical dynamics in isolated molecule^^^*^^ and investigations of dynamical processes in the condensed phase.*03
References and Notes (1) The Physics and Chemistry of Small Clusters;Jena, P., Rao, B. K., Khanna, N., Eds.; Plenum: New York, 1986. (2) The Chemical Physics of Atomic and Molecular ClustersProceedings of the Enrico Fermi International Summer Schools; Scolcs,G., Ed.; Elsevier Science: Amsterdam, 1990. ( 3 ) Elemental and Molecular Clusters; Benedek, G.,Martin, T. P., Pacchioni, G..Ed.;Springer-Verlag: Berlin, 1988. (4) Structure and Dynamics of Weakly Bound Molecular Complexes; Weber, A., Ed.;Reidel: Dordrecht, 1987. (J) Atomic and Molecular Clusters; Bernstein, E., Ed.; Elsevier: Amsterdam, 1990. (6) Dynamics of Polyatomic van der Waals Complexes; Halberstadt, N., Janda, K. C., Ed.;Plenum: New York, 1990. (7) Castleman, A. W.;K-, R. G. Acc. Chem. Phys. 1986, 19,413. (8) Naaman. R.Adu. Chem. Phys. 1988, 70, 18 1. (9) Janda, K. C.Adu. Chem. Phys. 1985, 60,201. (10) Alimi, R.;Gerber,R. B. Phys. Rev. Lett. 1990, 64, 1453. (1 1) Garcia-Vela,A.; Gerber,R. B.; Valentini, J. J. J . Chem. Phys. 1992, 97,3297. (12) Goyal, S.; Schutt, D. L.; Scolcs, G.Acc. Chem. Res. 1993,26, 123. (13) Hu,X.-C.; Martens, C. C. J . Chem. Phys. 1992, 97,8805.
Polyatomic Molecules in Rare-Gas Clusters (14) Hu. X.-C.; Hase, W. L. Z . Phys. D 1992,25, 57.
(15) Hunvitz, Y.;Rudich, Y.; Naaman, R.; Gerber, R. B. J . Chem. Phys. 1993,98,2941. (16) Potter, E. D.; Liu, Q.;Zewail, A. H. Chem. Phys. Lett. 1992,200,
605. (17) Schutz, M.; Wulfert, S.;Leutwyler, S.Z . Phys. D 1991, 20,247. (18) Shalev. E.; Jortner, J. Chem. Phys. Lert. 1991, 178,31. (19) Syage, J. A. In Ultrafast Specrroscopy in ChemicalSystems; Simon, J. D., Ed.; Kluwer Academic Publishers: Dordrecht, in press. (20) Zhao, X. G.; Tucker, S.C.; Truhlar, D. G. J. Am. Chem. Soc. 1991, 113,826. (21) Leutwyler, S.;Bosiger, J. Faraday Discuss. Chem. Soc. 1988,86, 225. (22) Alexander, M. L.; Levinger, N. E.; Johnson, M. A.; Ray, D.; Lineberger, W. C. J . Chem. Phys. 1988, 88, 6200. (23) Amar, F. G.; Berne, B. J. J . Phys. Chem. 1984,88,6720. (24) Burke, M. L.; Klemperer, W. J. Chem. Phys. 1993,98, 1797. (25) Fei, S.;Zheng, X.; Heaven, C.; Tellinghuisen, J. J . Chem. Phys. 1992, 97,6057. (26) Hu,X.-C.; Hase, W. L. J . Phys. Chem. 1992,96,7535. (27) Noorbatcha, I.; Raff, L. M.; Thompson, D. L. J. Chem. Phys. 1984, 81,5658. (28) Otto, B.; Schroeder, J.; Troe, J. J . Chem. Phys. 1984,81, 202. (29) Papanikolas, J. M.; Gord, J. R.; Levinger, N. E.; Ray, D.; Vorsa, V.; Lineberger, W. C. J. Phys. Chem. 1991, 95,8028. (30) Papanikolas, J. M.; Vorsa, V.; Nadal, M. E.; Campagnola, P. J.; Gord, J. R.; Lineberger, W. C. J . Chem. Phys. 1992,97,7002. (31) Ray, D.; Levinger, N. E.; Papanikolas, J. M.; Lineberger, W. C. J . Chem. Phys. 1989,91,6533. (32) Valentini, J. J.; Cross, J. B. J . Chem. Phys. 1982,77, 572. (33) Beck, T. L.; Berry, R. S.J . Chem. Phys. 1988,88, 3910. (34) Berry, R. S.;Beck, T. L.; Davis, H. L.; Jellinek, J. Adv. Chem. Phys. 1988, 70,75. (35) Berry, R. S.J. Chem. SOC.,Faraday Trans. 1990,86,2343. (36) Bosinger, J.; Leutwyler, S.Phys. Rev. Lett. 1987,59, 1985. (37) Bixon, M.; Jortner, J. J . Chem. Phys. 1989,91, 1631. (38) Cheng, H.-P.; Li, X.; Whetten, R. L.; Berry, R. S . Phys. Rev. A 1992, 46,791. (39) Hahn, M. Y.;Whetten, R. L. Phys. Rev. Lett. 1988,61,1190. (40) Jellinek, J.; Beck, T. L.; Berry, R. S.J . Chem. Phys. 1986,84,2783. (41) Kmetic, M.A.; Le Roy, R. J. J. Chem. Phys. 1991, 95,6271. (42) Leutwyler, S.;Furlan, A.; Knochenmuss, R.; Schutz, M.; Troxler, T.; Wulfert, S.2.Phys. D 1991, 20, 209. (43) Torchet, G. 2.Phys. D 1991, 20, 251. (44) Farges, J.;de Faraudy, M. F.; Raoult, B.;Torchet, G. J. Chem. Phys. 1973,59,3454. (45) Farges, J.; de Faraudy, M. F.; Raoult, B.;Torchet, G. J . Chem. Phys. 1986,84, 3491. (46) Wales, D. J.; Berry, R. S . J. Chem. Phys 1990, 92, 4283. (47) Adams. J. E.; Stratt. R. M. J. Chem. Phvs. 1990. 93, 1358, (48) Beck, T. L.; Marchioro 11, T. L. J . C h e k Phys. 1990, 93, 1347. (49) de Pujo, P.; Mestdagh, J.-M.; Visticot, J.-P.; Cuvellier, J.; Meynadier, P.; Sublemontier. 0.;Lallement, A.; Berlande, J. Z . Phys. D. 1993,25,357. (50) Martens, C. C. Phys. Rev. A 1992,45,6914. (51) Adams, J. E.; Stratt, R. M. J. Chem. Phys. 1990, 93, 1632. (52) Cuvellier, J.; Mestdagh, J. M.; Meynadier, P.; de Pujo, P.; Sublemontier, 0.;Visticot, J. P.; Berlande, J. Chem. Phys. Lett. 1991, 176,325. (53) Finney, L. M.; Martens, C. C. J . Phys. Chem. 1992,96, 10626. (54) Garcia-Vela, A.; Gerber, R. B.; Valentini, J. J. Chem. Phys. Lett. 1991, 186,223. ( 5 5 ) Honma, K.; Kajimoto, 0. J . Chem. Phys. 1984,81, 3344. (56) Hu. X.-C.: Martens, C. C. J . Chem. Phvs. 1993,98,8551. (57) Hu; X.; Hase, W. L. J . Chem. Phys. 1993,98,7826. (58) Kaukonen, H.-P.; Landman, U.; Cleveland, C. L. J . Chem. Phys. 1991, 95,4997. (59) Koehler, J.; Deal, R. M.; Hase, W. L.; Lu, D. J . Cluster Sci. 1990, I, 335. (60) Lallement, A.; Cuvellier, J.; Mestdagh, J. M.; Meynadier, P.; de Pujo, P.; Sublemontier, 0.;Visticot, J. P.; Berlande, J.; Biquard, X. Chem. Phys. Lett. 1992,189, 182. (61) Nieman, J.; Naaman, R. Chem. Phys. 1984,90,407.
The Journal of Physical Chemistry, Vol. 97, No. 51, 1993 13483 (62) Nieman, J.; Shwartz, J.; Naaman, R. Z . Phys. D 1986,1, 231. (63) Nieman, J.; Naaman, R. J. Chem. Phys. 1986,84,3825. (64) Shin, S.K.; Chen, Y.;Nickolaisen. S.;Sharpe, S.W.; Beaudet, R. A.; Wittig, C. Adu. Photochem. 1991, 16, 249. (65) Tardiff, J.; Deal, R. M.; Hase, W. L.; Lu, D. J . Cluster Chem. 1990, 1, 335. (66) Visticot, J. P.; Mestdagh, J. M.; Alcaraz, C.; Cuvellier, J.; Berlande, J. J . Chem. Phys. 1988,88,3081. (67) Wittig, C.;Sharpe,S.; Beaudet, R.A. Acc. Chem. Res. 1988,21,341. (68) Lurge Finite System; Jortner, J., Pullman, A., Pullman, B., Eds.; D. Reidel: Dordrecht, 1987. (69) Jortner, J. 2.Phys. D 1992,24, 247. (70) Jortner, J.; Even, U.; Ben-Horin, N.; Scharf, D.; et al. 2.Phys. D 1989,12, 167. (71) Hoare. M. R. Ado. Chem. Phvs. 1979. 40. 49. (72) Farges', J.; de Faraudy, M. F.; Raoult, B:;Torchet, G. J . Chem. Phys. 1983. 78. 5067. ~. (f3) -Farges, J.;de Faraudy, M. F.; Raoult, B.;Torchet,G. In Lorge Finite Systems; Jortner, J., Pullman, A., Pullman, B., Eds.; Reidel: Dordrecht, 1987; p 113. (74) Torchet,G.;Farges, J.;deFeraudy,M. F.;Raoult,B.In Thechemistry Physics of Atomic and Molecular Clusters. Proceedings of ihe International School of Physics 'Enrico Fermi", Course CVIk Scoles, G., Ed.; North Holland: Amsterdam, 1990; p 513. (75) Berry, R. S.2.Phys. D 1989,12, 161. (76) Bemy, R. S.In Lurge FinireSystems;Jortner, J., Pullman, A., Pullman, B.. Eds.: Reidel: Dordrecht. 1987: D 135. .(77)-Goyal,S.;Robinson; G. N:;'Schutt, D. L.; Sales, G. J. Phys. Chem. 1991, 95,4186. (78) Jortner, J.; Ben-Horin, N. J. Chem. Phys. 1993,98,9346. (79) Jortner, J.; Levine, R. D.; Rice, S. A. Adu. Chem. Phys. 1988,70, I.
(80) Perera, L.; Amar, G. J. Chem. Phys. 1989,90,7454. (81) Amar, F. G.; Perera, L. Z . Phys. D 1991, 20, 173. (82) Perera, L.; Amar, F. G. J . Chem. Phys. 1990, 93,4884. (83) Gough, T. E.; Mengel, M.; Rowntree, P. A.; Scoles, G. J. Chem. Phys. 1985,83,4958. (84) Levandier, D. J.; McCombie, J.; Pursel, R.; Scoles,G. J . Chem. Phys. 1987,86,7239. (85) Levandier, D. J.; Mengel, M.; Pursel, R.; McCombie, J.; Scoles,G. 2.Phys. D-At., Mol. Clusters 1988,10, 337. (86) Levandier, D. J.; Goyal, S.;McCombie, J.; Pate, B.; Scoles, G. J . Chem. Soc., Faraday Trans. 1990,86,2361. (87) Crim, F. F. Annu. Rev. Phys. Chem. 1984,35, 657. 188) Crim. F. F. Science 1990.249. 1387. (89) Butler, L. J.;Ticich. T. M:; Likar, D. M.; Crim, F. F. J . Chem. Phys. 1986,85,2331. (90) Ticich, T. M.; Rizzo, T. R.; Dubal, H.-R.; Crim, F. F. J . Chem. Phys. 1986,84,1508. (91) Luo, X.; Rieger, P. T.; Perry, D. S.;Rizzo, T. R. J. Chem. Phys. 1988. 89,4448. (92) Luo. X.: Rizzo. T. R. J . Chem. Phvs. 1990. 93. 8620. (93j Schkrer; N. F.;'Doany, F. E.; ZewGl, A. H.f Perry, J. W. J . Chem. Phys. 1986.84, 1932. (94) Scherer, N. F.; Zewail, A. H. J . Chem. Phys. 1987,87, 97. (95) SumDter, B. G.; Thompson, D. L. Chem. Phvs. Lett. 1988,153.243. (96) Sumpter, B. G.; Thompson, D. L. J. Chem. Phys. 1985, 82, 4557. (97) Sumpter, B. G.; Thompson, D. L. J. Chem. Phys. 1987,86, 2805. (98) Uzer, T.; Hynes, J. T.; Reinhardt, W. P. Chem. Phys. Lett. 1985, 117,600. (99) Uzer, T.; Hynes, J. T.; Reinhardt, W. P. J . Chem. Phys. 1986,85, 5791. (100) Uzer, T.; MacDonald, B. D.; Guan, Y.;Thompson, D. L. Chem. Phvs. Lett. 1988. 152. 405. 1101) Getino,' C.; Sumpter, B. G.; Santamaria, J.; Ezra, G. S.J. Phys. Chem. 1989,93, 3877. (102) Getino, C.; Sumpter, B. G.; Santamaria, J. Chem. Phys. 1990,145, I.
(103) Li, Y. S.;Whitnell, R. M.; Wilson, K.R.; Levine, R. D. J . Phys. Chem. 1993,97,3647. (104) Gear, C. W. SIAM J. Numer. Anal. 1964,28,69.